I was just asking how an online advertising service, especially a small startup, handle such things.
Nobody who says "the more accurately you judge my efforts, the worse off I'll be" is worth keeping around.
Right now, it is relatively easy for a company to sign up for an online advertising service with some minimal guaranteed ad views. From that companies can project future earnings based on their click through and purchase rates, even if their ads are terrible, and nobody except for 0.1% click, they at least have a base line for money coming in that is fairly reliable. This system, however bad for the end users experience, is extremely helpful for the businesses planning and forecasting. Now, the CFO can say, 'We're going to have at least this much money!', and Net Ops groups can say, 'Well that means we'll probably need this much bandwidth!', and the entire business can some-what get along.
Then, Adpinion steps in. Suddenly, Advertising/Marketing can't predict how much, or even if, they'll be able to meet previous projections, projections are useless at this point. Now the CFO can't say if the company will be able to meet market predictions, and the Net Ops guys will be in the dark about if they'll suddenly be crushed by users' bandwidth requests or that they've over allocated bandwidth and they're now stuck with some fat pipes that are just getting spam messages. Well, this is all a little dire, but the point is that Adpinion can create contentions in a business.
Ah! But here is what I wanted to follow through on - even if Adpinion can create unpredictability in long running advertising strategies, the real win here would be for a one-two punch with traditional online advertising venues. A business could market test half their advertising revenue, find what works and then send it through traditional online ad streams. Now that business can up their average click rate by quickly market testing ads, and they can make their advertising team hungry and on-edge about what how well that day/week/months' ad scheme performed. And this is where Adpinion wins, for me. From an overall business perspective, outside of a localized ad/marketing department, it would provide such a benefit to any company that they would be senseless not to use it.
The issue of advertisements getting shown less because everyone dislikes them is something we're looking at closely. We focus a lot on individual user preferences, so if you dislike an ad hopefully there is someone out there who likes it and they will get shown the ad more to keep the number of impressions consistent. However, there will be cases where an ad gets less impressions because the majority of people dislike it. We feel that serving these ads as much as the ads that everyone likes creates an inefficiency in the cost per click model: if you're a website publisher and there are two ads that pay the same per click, you're not making as much off the less popular ad. We're exploring ways to incorporate ad ratings into the cost structure in a way that gives advertisers incentive to make well-liked ads.
Also: guaranteeing a minimum viewership/maximum price for cost estimation is definitely something we're aware of, and we have some ideas on how to achieve that. We'll reveal them when we're done ; ).